ADIDRUS: Using Deception to Protect and Defend UAVs

Unmanned vehicles are becoming an increasingly valuable tool within the Department of Defense, and related technologies are beginning to appear in commercial systems. As their utilization grows, these unmanned systems will become a significant target for cyber attack. Indeed, existing Defense Department unmanned aerial vehicles (UAVs) have already been subjected to attack by various adversaries. One of the primary reasons that our cyber adversaries have had such success is due to the use of deception in their operations. For instance, among APT adversaries, their standard means of gaining an initial foothold is through social engineering: a target is fooled into giving away personal information that the adversary exploits to gain access to a target resource. Deception plays a role in virtually any cyberattack simply for the reason that it works, and the historical evidence bears this out. The very effectiveness of deception suggests that it should be considered as a means to fight back against cyber adversaries. Our response to the challenge of applying deception to the domain of cyberattacks on UAVs is the creation of a platform called ADIDRUS --'Adversarial Dectection, Inference, and Deceptive Response for Unmanned Systems'. ADIDRUS is designed to be a platform-agnostic, software-based solution that incorporates all three of the tactics above into a system that hardens UAVs against attacks through the use of adversarial reasoning models and deceptive tactics turned against the adversary.